public inbox for gentoo-science@lists.gentoo.org
 help / color / mirror / Atom feed
* [gentoo-science] [sage-on-gentoo] updates
@ 2010-05-21 10:40 François Bissey
  2010-05-21 20:55 ` Steven Trogdon
  0 siblings, 1 reply; 3+ messages in thread
From: François Bissey @ 2010-05-21 10:40 UTC (permalink / raw
  To: gentoo-science

Hi all,

sage-4.4.2 has just been released. I am guessing that Christopher
may be preparing for exams in June so I will probably do the bump.
I have already bumped sage-doc.

Other matters:
-The situation on amd64 is still bad. There has been a silent (no revbump)
update to cython to deal with a strict-aliasing problem:
http://bugs.gentoo.org/show_bug.cgi?id=294585
I don't have the hardware so I cannot check if it has an influence on our
problems but it is worth a try.

-python-2.6.5-r2 is going stable. sage is of course still shipping 2.6.4.
I have already put a "pickled" version of python-2.6.5-r2 in the overlay 
(2.6.5-r99). Apart from:  
http://github.com/cschwan/sage-on-gentoo/issues#issue/1
which needs to be fixed in sage (patch/suggestions welcome) I don't expect
any problems with that.

-mpir-2.0.0 entered the overlay, sage is actually still shipping 1.2.2.
Report problems as usual.

-I have tracked 2 other failing tests which are now working:
http://github.com/cschwan/sage-on-gentoo/issues#issue/8
http://github.com/cschwan/sage-on-gentoo/issues#issue/9

-I am wondering if issues 3 and 6 are down to -ffast-math in one way
or another. If you have any similar issues please post.

-still no clue on gap's behaviour. Considering the amount of aliasing problem
you may as well try to add -fno-strict-aliasing to your CFLAGS for building 
gap and see if that improve things.

-I have imported sympy-0.6.6 in the overlay as it has left the portage tree
and sage currently doesn't work with 0.6.7. We should investigate doing
some porting work. Porting work to scipy-0.7.2 would be a good idea too.

Francois



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [gentoo-science] [sage-on-gentoo] updates
  2010-05-21 10:40 [gentoo-science] [sage-on-gentoo] updates François Bissey
@ 2010-05-21 20:55 ` Steven Trogdon
  2010-05-22 10:33   ` Christopher Schwan
  0 siblings, 1 reply; 3+ messages in thread
From: Steven Trogdon @ 2010-05-21 20:55 UTC (permalink / raw
  To: gentoo-science

[-- Attachment #1: Type: text/plain, Size: 997 bytes --]

On   05/21/10 - 05:40:38, François Bissey wrote:
> Hi all,
> 
> sage-4.4.2 has just been released. I am guessing that Christopher
> may be preparing for exams in June so I will probably do the bump.
> I have already bumped sage-doc.
> 
> Other matters:
> -The situation on amd64 is still bad. There has been a silent (no  
> revbump)
> update to cython to deal with a strict-aliasing problem:
> http://bugs.gentoo.org/show_bug.cgi?id=294585
> I don't have the hardware so I cannot check if it has an influence on  
> our
> problems but it is worth a try.
> 

Just for the record, I've tried sage-4.4.1 and sage-4.4.2 with a  
rebuild of cython and I still have the random "corrupted double-linked  
list" failures on my amd64. Has anyone had any success using valgrind  
to debug this. Is it possible to even use portage valgrind to debug  
Sage? I have 3.5.0 installed, built with FEATURES=splitdebug and Sage  
seems to cleverly avoid all attempts at using it.

Steve


[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [gentoo-science] [sage-on-gentoo] updates
  2010-05-21 20:55 ` Steven Trogdon
@ 2010-05-22 10:33   ` Christopher Schwan
  0 siblings, 0 replies; 3+ messages in thread
From: Christopher Schwan @ 2010-05-22 10:33 UTC (permalink / raw
  To: gentoo-science

Hi,

On Friday 21 May 2010 22:55:03 Steven Trogdon wrote:
> On   05/21/10 - 05:40:38, François Bissey wrote:
> > Hi all,
> > 
> > sage-4.4.2 has just been released. I am guessing that Christopher
> > may be preparing for exams in June so I will probably do the bump.
> > I have already bumped sage-doc.
> > 
> > Other matters:
> > -The situation on amd64 is still bad. There has been a silent (no
> > revbump)
> > update to cython to deal with a strict-aliasing problem:
> > http://bugs.gentoo.org/show_bug.cgi?id=294585
> > I don't have the hardware so I cannot check if it has an influence on
> > our
> > problems but it is worth a try.
> 
> Just for the record, I've tried sage-4.4.1 and sage-4.4.2 with a
> rebuild of cython and I still have the random "corrupted double-linked
> list" failures on my amd64. Has anyone had any success using valgrind
> to debug this. Is it possible to even use portage valgrind to debug
> Sage? I have 3.5.0 installed, built with FEATURES=splitdebug and Sage
> seems to cleverly avoid all attempts at using it.

I think you already saw this thread (if not this is what you are looking for):

http://groups.google.com/group/sage-
devel/browse_thread/thread/cd8005be927fc896

To use valgrind on amd64 follow these steps:

- emerge valgrind-3.5.0 (previous versions did not work)
- valgrind prints out a message to reinstall libc (or was it glibc ?) with 
FEATURES=splitdebug - thats defintely needed
- edit sage-valgrind in Sage's local/bin directory: remove "--
suppressions=..." option or point to this file (see thread above) and add --
trace-children=yes
- start sage with "sage -valgrind"

> 
> Steve

Christopher



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2010-05-22 10:33 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-05-21 10:40 [gentoo-science] [sage-on-gentoo] updates François Bissey
2010-05-21 20:55 ` Steven Trogdon
2010-05-22 10:33   ` Christopher Schwan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox